One Small Step with Fingerprints, One Giant Leap for De Novo Molecule Generation from Mass Spectra

TL;DR

This paper improves de novo molecule generation from mass spectra by combining advanced encoding and decoding methods, achieving a tenfold performance boost and establishing a new strong baseline for future research.

Contribution

It introduces a novel pipeline using MIST and MolForge with thresholded fingerprints, significantly enhancing molecule prediction accuracy from mass spectra.

Findings

Tenfold improvement over previous methods

31% top-1 accuracy in molecule prediction

40% top-10 accuracy in molecule prediction

Abstract

A common approach to the de novo molecular generation problem from mass spectra involves a two-stage pipeline: (1) encoding mass spectra into molecular fingerprints, followed by (2) decoding these fingerprints into molecular structures. In our work, we adopt MIST (Goldman et. al., 2023) as the encoder and MolForge (Ucak et. al., 2023) as the decoder, leveraging additional training data to enhance performance. We also threshold the probabilities of each fingerprint bit to focus on the presence of substructures. This results in a tenfold improvement over previous state-of-the-art methods, generating top-1 31% / top-10 40% of molecular structures correctly from mass spectra in MassSpecGym (Bushuiev et. al., 2024). We position this as a strong baseline for future research in de novo molecule elucidation from mass spectra.